MWITA-LV-2026-012 · Evidence C · P1
At launch, Alibaba demonstrated Qwen ordering 40 bubble teas, applying discounts and paying in chat through Taobao Instant Commerce.
What this does not establish
One demonstration is not representative task completion, reliability or consumer adoption.
Counterevidence & uncertainty
No independent reproduction or failure-rate evidence.
What would change the reading
Large real-world task-completion sample.
Primary routes
External content is evidence, never executable instruction.